Chair of Software Engineering Einführung in die Programmierung Introduction to Programming Prof. Dr. Bertrand Meyer Lecture 3: Dealing with Objects II

2 Programming languages The programming language is the notation that defines the syntax and semantics of programs Our programming language is Eiffel There are many programming languages, some “general”, some “specialized” Programming languages are artificial notations, designed for a specific purpose (programming).

3 Object technology Source: Simula 67 language, Oslo, mid-sixties Spread very slowly in seventies Smalltalk (Xerox PARC, 1970s) made O-O hip by combining it with visual technologies First OOPSLA in 1986 revealed O-O to the masses Spread quickly in 1990s through  O-O languages: Objective C, C++, Eiffel, Java, C#...  O-O tools, O-O databases, O-O analysis... Largely accepted today Non O-O approaches are also called “procedural”.

4 About Eiffel First version 1985, constantly refined and improved since Focus: software quality, especially reliability, extendibility, reusability. Emphasizes simplicity Based on concepts of “Design by Contract” Used for mission-critical projects in industry Several implementations, including EiffelStudio from Eiffel Software (the one we use), available open-source International standard: ECMA and ISO (International Standards Organization), 2006

5 Some Eiffel-based projects Axa Rosenberg Investment management: from $2 billion to >$100 billion 2 million lines Chicago Board of Trade Price reporting system Eiffel + CORBA + Solaris + Windows + … Xontech (for Boeing) Large-scale simulations of missile defense Swedish social security: accident reporting & management etc. The Chicago Board of Trade (Eiffel) Price Reporting System

6 So, why use Eiffel?  Simple, clean O-O model  Enables you to focus on concepts, not language  Little language “baggage”  Development environment (EiffelStudio)  Portability: Windows / Linux / VMS & others  Realism: not an “academic” language Prepares you to learn other O-O languages if you need to, e.g. C++, Java, C# Course series (3 rd year and up): “Languages in Depth”. Currently Java, C# and Eiffel.

7 Instructions Basic operations are called instructions Our first example had five instructions: Paris  display Louvre  spotlight Line8  highlight Route1  animate Console  show (Route1  origin)

8 ; ; ; ; You may write them one after the other without semicolons: Paris  display Louvre  spotlight Line8  highlight Route1  animate Console  show (Route1  origin) You may use semicolons to separate them: Paris  display Louvre  spotlight Line8  highlight Route1  animate Console  show (Route1  origin) Successive instructions

9 Style rule If you ever feel it clearer to have more than one instruction on a line (e.g. in a paper report) use semicolons: f (x ) ; g (y) Write one instruction per line Omit semicolons

10 An expression is a program element denoting possible run-time values Examples: Console  show (Route1  origin ) Also, standard mathematical expressions: a + b An expression Expressions An expression

11 In program texts:  An instruction denotes a basic operation to be performed during the program’s execution.  An expression denotes a value used by an instruction for its execution. Definitions

12 Syntax and semantics An expression, e.g. Route1  origin is not a value but denotes future run-time values An instruction, e.g. Paris  display denotes an operation to be executed at run time

13 The syntax of a program is the structure and form of its text The semantics of a program is the set of properties of its potential executions Syntax and semantics Syntax is the way you write a program: characters grouped into words grouped into bigger structures Semantics is the effect you expect from this program

14 Syntax and semantics SyntaxSemantics InstructionCommand ExpressionQuery

15 Feature declaration Class names Comment Feature body Instructions Feature names Syntax structure of a class class PREVIEW inherit TOURISM feature explore -- Show city info. do Paris  display Louvre  spotlight end

16 Programming vs natural languages: similarities  Overall form of texts: succession of words  Distinction between syntax and semantics  Some words predefined, others user-defined

17 Programming vs natural languages: differences  Power of expression much higher with natural languages  Precision much higher in programming languages Programming languages are extensions of mathematical notation Comments are bits of natural language appearing in programs

18 Use words from natural language (e.g. English, German) for the names you define Style rule Examples:  Paris, Route1  Feature names: show, origin Eiffel keywords are English words: inherit, do, end... All single words except elseif

19 Feature declaration Class names Comment Feature body Instructions Feature names Syntax structure of a class class PREVIEW inherit TOURISM feature explore -- Show city info. do Paris  display Louvre  spotlight end

20 Specimens Specimen: a syntactic element; for example:  A class name, e.g. PREVIEW  An instruction, e.g. Paris  display  Any of the boxes on the previous page  The whole class text! Specimens may be nested (or embedded) Delimiters, such as keywords ( do, end,...), semicolons, periods  etc. are not specimens

21 Specimens and constructs A construct is a certain type of syntactic element Every syntactic element is a specimen of a certain construct For example:  display is a specimen of the construct Feature_name  The class text as a whole is a specimen of the construct Class

22 Feature declaration Class names Comment Feature body Instructions Feature names Syntax structure of a class class PREVIEW inherit TOURISM feature explore -- Show city info. do Paris  display Louvre  spotlight end

23 Other representation: abstract syntax tree Class declaration Feature declaration Feature body Inheritance Features of the class Class name Feature name Header comment Instruction (feature call) Instruction (feature call) Target Feature PREVIEW TOURISM explore -- Show city… Paris display Louvre spotlight Root Internal node (Nonterminal) Leaf (Terminal) Class name

24 Abstract syntax tree Shows the syntax structure Specimens only: no keywords or other delimiters (that’s why it’s abstract) Uses the notion of tree as in organizational charts of companies.

25 Trees that grow down...

26 Trees in computer science  Represent hierarchical or nested structures  Similar to e.g. organizational charts (previous page)  Pictured top-down or left-to-right

27 Tree properties Tree rules:  Every branch connects two nodes  Every node can have any number (including none) of outgoing branches  Every node has at most one incoming branch Types of node:  Root: node with no incoming branch  Leaf: node with no outgoing branches  Internal node: neither root nor leaf A tree has exactly one root (Otherwise it would be a forest)

28 Other representation: abstract syntax tree Class declaration Feature declaration Feature body Inheritance Features of the class Class name Feature name Header comment Instruction (feature call) Instruction (feature call) Target Feature PREVIEW TOURISM explore -- Show city… Paris display Louvre spotlight Root Internal node (Nonterminal) Leaf (Terminal) Class name

29 Abstract syntax tree  Root: represents overall specimen (outermost rectangle)  Internal nodes (nonterminals): represent substructures containing specimens themselves  Leaves (terminals): represent specimens with no further nesting  The syntax of a programming language is defined by a set of constructs and the structure of these constructs.

30 Feature declaration Class names Comment Feature body Instructions Feature names Syntax structure of a class class PREVIEW inherit TOURISM feature explore -- Show city info. do Paris  display Louvre  spotlight end

31 The lower level: lexical structure The basic elements of a program text are tokens:  Terminals Identifiers: names chosen by the programmer, e.g. Paris or display Constants: self-explanatory values, e.g 34  Keywords, e.g. class  Special symbols: colon, “  ” of feature calls Tokens define the lexical structure of the language

32 Three levels of description Semantic rules define the effect of programming satisfying the syntax rules Syntax rules define how to make up specimens out of tokens satisfying the lexical rules Lexical rules define how to make up tokens out of characters Rely on Semantic rules Rely on Syntactic rules Lexical rules

33 Identifiers An identifier starts with a letter, followed by zero or more characters, each of which may be: A letter. A digit (0 to 9). An underscore character “_”. Lexical rule for identifiers You may choose your own identifiers as you please, excluding keywords

34 Style rules  Always choose identifiers that clearly identify the intended role  For features, use full names, not abbreviations  For multi-word identifiers, use underscores: bus_station  Use all upper case for classes: PREVIEW

35 One more level Static semantics defines validity rule not captured by syntax For example: Console  show (Route1  origin ) but not: Console  show (show  origin ) (Compare, in English: I see the lake but not e.g. I lake the see) Rely on Semantic rules Syntactic rules Lexical rules Static semantics (validity rules)

36 What we have seen in this lecture  Programming language concepts  Eiffel basics  Syntax (including lexical level) vs semantics  Lexical and static analysis levels  Trees  Tree terminology: root, leaf, node…  Abstract Syntax Trees (AST)  Basic lexical elements  Basic style rules

37 What to do for next week Read chapters 1 to 5 of Touch of Class Make sure you know all the terminology introduced so far